The overall goal of this project is to develop promising drug candidate leads using a new drug design paradigm based on initial three-dimensional RNA-structure-based computational screening of about 255,000 commercially available compounds for binding to selected RNA targets using the three-dimensional structure of portions of the HIV-1 genome. Specifically, the transactivation response element (TAR) and the dimer linkage site (DLS) are proposed as targets. The latter is a novel target for drug design. Good computational "hits" will be tested for inhibition in functional assays. The best of these will be utilized in NMR studies to verify the mode of binding to the RNA target. The best candidates will be water-soluble, nonpeptide, nonnucleotide organic compounds generally with molecular weight less than 500 daltons, but it may be necessary to consider slightly larger compounds as leads. The NMR structural studies of complexes formed with potential leads and their RNA targets should yield insight into features governing affinity and specificity. Promising leads will be selected for further development via computational and experimental combinatorial chemistry. Some further development of the methodology is proposed.